SPQ SNAPS
Should I Stay or Should I Go? Reciprocity, Negotiation, and the Choice of Structurally
Disadvantaged Actors to Remain in Networks
Scott V. Savage and Zachary L. Sommer
Abstract
We consider how the two forms of direct exchange influence whether structurally disadvantaged
actors choose to stay in the micro-structures that disadvantage them. We posit that a history of
reciprocal, as opposed to negotiated, exchange should increase the likelihood of an exit
opportunity resulting in disadvantaged actors defining themselves as members of a group rather
than as unique individuals in a network, and that this should increase the likelihood of actors
choosing to remain in reciprocal exchange networks. Findings from two laboratory experiments
generally support our argument that for disadvantaged actors, psychological group formation
mediates the relationship between exchange form and staying in networks.
Keywords
social exchange, self-categorization theory, network dynamics, group formation
Why might someone remain committed to a romantic partnership when the possibility exists of
finding a more meaningful relationship elsewhere? Similarly, why might people bypass
alternative job opportunities and stay with the companies that exploit them? Questions like these
motivate research on the stability of network exchange structures that privilege some actors and
disadvantage others (e.g., Cook and Gillmore 1984; Leik 1992; Willer and Willer 2000; Dogan
et al. 2009; Rand, Arbesman, and Christakis 2011). For the most part, this research has focused
on how adding and/or deleting linkages to others in a network can improve or worsen one’s
relative power. Willer and Willer (2002), for example, concluded that because adding links
generally benefits those in low power positions, rational actors in low power positions should try
to add linkages in an attempt to balance power in the network. To counteract this and maintain
power, the power advantaged should respond by limiting the ability of others to seek out new
connections.
One way of accomplishing this is by creating a ‘with us or against us’ scenario, whereby
individuals must choose between existing and alternative sets of network ties. Although extreme,
such situations do occur. Employers often demand the full commitment of employees, forcing
them to choose between the current employer and alternative employers (Hirschman 1970; Kmec
2007). Children sometimes do something similar when constructing friendship groups. Adler and
Adler (1995), for example, showed that joining a more popular friendship network sometimes
means cutting ties with one’s current set of friends. That it is possible to sever ties with one
network in order to join a new one, however, does not address why those disadvantaged by the
structure of the network might choose to remain in it.
One possible explanation hinges on gift giving and the norm of reciprocity. Ongoing
research by social exchange theorists finds that the structure of reciprocal exchange can serve as
2
an integrating force (e.g., Molm, Collett, and Schaefer 2007; Molm, Whitham, and Melamed
2012) that can motivate the development of psychological contracts or beliefs in the mutual
obligation of self and other (Rousseau 1989; Rousseau and Tijoriwala 1998). Building on this
insight, we investigate how different exchange histories shape whether disadvantaged actors
choose to stay in the very networks that disadvantage them. We theorize 1) that the exit
opportunity is more likely to result in disadvantaged actors coming to view their network as a
group if there has been a history of reciprocal, as opposed to negotiated, exchange and 2) that
this psychological group formation should affect the exit choice.
THEORY
Social Exchange
The social exchange perspective views human interaction as a series of resource exchanges
between actors (Homans 1974). Resources in social exchange are any capability or object that is
held by one actor and valued by another. Exchange relationships form when the resources to be
had from interdependently exchanging with others outweigh independent action (Thibaut and
Kelley 1959).
Resource exchanges can take various forms. When the outcomes of exchange depend on
the behaviors of actors who exchange with one another, an exchange relation is said to be direct,
and there are two forms of direct exchange relations: negotiated and reciprocal. Negotiated
exchanges are those where actors bargain over the terms of the exchange and engage in an
exchange transaction only after those terms have been agreed on. Negotiated exchanges are
characteristic of traditional markets, as exemplified by sports teams bargaining with one another
over the terms of a trade and players bargaining with sports teams over the terms of a contract
(see Lawler, Thye, and Yoon (2009) for an example).
3
Reciprocal exchanges are those where actors independently initiate an exchange by
providing a benefit to another, e.g., swapping birthday presents. Lack of negotiation and binding
agreements means that benefits flow unilaterally and that there is no assurance that giving will
result in another’s reciprocation (e.g., Molm 1994; Molm et al. 2007; Cheshire, Gerbasi, and
Cook 2010).
Although both forms of exchange occur directly between two actors, these actors may be
embedded in more complex networks of exchange relations. Such exchange structures consist of
the connections between a set of exchange actors and shape the mutual dependence of actors.
Negatively connected network exchange structures are those where exchange in one relation
precludes exchange in another on any given exchange opportunity. For example, if an individual
wants to buy a car and he chooses to buy it from a certain dealer, this precludes buying the car
from all other dealers. Power imbalances in such networks arise due to inequalities in
dependence. When network alternatives result in one actor being less dependent on another actor
for resources than the other actor is on her, a power imbalanced exchange relation exists, with
the less dependent actor experiencing a power advantage (Emerson 1972).
We consider why those disadvantaged by these structural conditions sometimes opt to
stay in them and whether this choice is influenced by the form of exchange. Key to our argument
is the mechanism of psychological group formation or the process by which someone comes to
view herself as part of a group of three or more actors characterized by increased cooperation
and reduced levels of self-interest (Hogg 1992; Lawler and Yoon 1996).
Form of Exchange, Group Formation, and Disadvantaged Actors Staying in Networks
The structure of exchange can have important consequences for the bonds that develop between
actors (e.g., Molm et al. 2007; Lawler, Thye, and Yoon 2008; Kuwabara 2011).
4
Molm and colleagues (2007), for example, find that actors in negotiated and reciprocal exchange
networks view the bonds that exist between exchange partners as distinct and isolated units, with
those embedded in reciprocal exchange networks expressing greater personal attractions for
particular exchange partners than those embedded in negotiated exchange networks (Molm et al.
2007). These actors, however, rarely describe their set of exchange relations as a “group, a team,
working together” (Molm et al. 2007: 234). This finding reinforces Hogg and Turner’s (1985)
claim that while personal attractions can serve as the basis for group formation, they are unlikely
to influence group formation if the relationships are idiosyncratic and personal. So even though
reciprocal exchange results in stronger bonds between pairs of exchange partners, i.e., personal
attractions, than negotiated exchange (Molm, Takahashi, and Peterson 2000; Molm, Collett, and
Schaefer 2006; Molm et al. 2007; Molm et al. 2012), actors in reciprocal exchange networks
appear no more likely to view their networks as groups.
We argue that the exit opportunity changes this by introducing into the situation a
categorical distinction that can serve as the basis for an in-group/out-group distinction. Selfcategorization theory assumes a social situation will motivate psychological group formation if
1) an individual’s past experiences and present desires prepare her to use a perceived categorical
distinction and 2) she thinks categorization minimizes intra-class differences relative to interclass differences in expected or stereotypical ways (Turner, Oakes, Haslam, and McGarty 1994;
Turner and Reynolds 2012).
A history of reciprocal exchange is more likely to satisfy these conditions for
disadvantaged actors than a history of negotiated exchange. Because disadvantaged actors
experience negotiation as more conflict-laden and less fair than reciprocity (Molm, Takahashi,
Peterson 2003; Molm, Collett, and Schaefer 2006), a history of negotiation should result in
5
disadvantaged actors being less prepared to use the categorical distinction as the basis for
psychological group formation. Moreover, because negotiation clashes with the friendship norms
of welfare and fairness (Kurtzberg and Medvec 1999), it should be easier for disadvantaged
actors to minimize intra-class differences and to view exchange partners as behaving in ‘groupy’
ways if they had previously engaged in reciprocal exchange. Thus, we hypothesize:
Hypothesis 1: The exit opportunity will be more likely to result in psychological group
formation for disadvantaged actors in reciprocal exchange networks than in negotiated
exchange networks.
We also suspect that perceptions of personal attraction will mediate this relationship.
How resources are exchanged across the two forms of direct exchange results in reciprocal
exchange being riskier, less conflict-laden, and more symbolically valuable than negotiated
exchange, and this results in greater feelings of personal attraction for specific exchange partners
in reciprocal exchange networks (Molm et al 2007). These differences in personal attraction,
which we define as perceptions of social unity with a specific exchange partner, become relevant
when actors become aware of the opportunity to join an alternative network insofar as they
provide “a cognitive criterion for common category membership” (Hogg and Turner 1985:61).
This is likely because positive personal attractions amplify perceptions of similarity (e.g.,
Backman and Secord 1962; Hogg and Turner 1985), thereby resulting in the minimization of
intra-class differences relative to inter-class differences (Turner and Reynolds 2012).
Consequently, we contend that following the exit opportunity, disadvantaged actors will rely on
their personal attractions as a basis for group formation and because personal attractions differ
across the two forms of direct exchange, personal attractions should mediate the relationship
between exchange form and psychological group formation.
6
Hypothesis 2: When disadvantaged actors become aware of the opportunity to join an
alternative network, their personal attractions for their respective exchange partners will
positively affect whether they view their existing network as a group.
Hypothesis 3: Personal attraction will mediate the relationship between exchange form
and psychological group formation.
That the exit opportunity should differentially affect psychological group formation
across the two forms of direct exchange matters, because how strongly an individual identifies
with a group should affect whether the actor stays in it (Hogg 1992). Thus, whether an actor
chooses to remain in an existing exchange network or not should depend on whether the actor
views the network as a group to which he or she belongs.
Hypothesis 4: Psychological group formation group will positively affect whether
disadvantaged actors choose to remain in an existing exchange network.
Hypothesis 5: Psychological group formation will mediate the relationship between the
form of exchange and staying in an existing network.
We test these hypotheses with data from two laboratory experiments.
EXPERIMENT 1
Experiment 1 involved undergraduate students earning money by exchanging with two
computer-simulated actors. The amount earned depended on participants’ exchanges with
partners.
Design and Participants
Social exchange theory assumes actors value the resources of exchange (Molm and Cook 1995).
Thus, we recruited undergraduate students from a large, public university based on their desire
for the resource in the experiment: money. The experiment manipulated the form of exchange
7
(reciprocal vs. negotiated exchange) and the information subjects had about the alternative
network (low vs. high information).1 Sixty-four undergraduate students were randomly assigned
to each of the experimental conditions created by crossing the form of exchange and the amount
of information so that there were 16 students per condition. 2 Equal numbers of men and women
were assigned to each condition.3
Procedures
Upon arriving, participants were directed to isolated rooms where they were told that instructions
for the experiment would appear on the computer terminal once all participants had arrived.
Instructions informed subjects that they would be interacting with other students and the amount
of money they earned would depend on these interactions. Unbeknownst to participants,
exchange partners were simulated actors designed to mimic real exchange behavior. Simulated
actors were used to ensure that subjects had comparable earnings across the forms of exchange.
The instructions also taught participants how to use their computers to make exchanges.
Participants were told that they would have two exchange partners who each had a common
alternative partner. A diagram of this exchange network appeared on subjects’ screens and can
be seen in Figure 1. Participants also learned they would be interacting with their partners over a
series of exchange opportunities and that for any one opportunity they would only be able to
exchange with one partner.4 Subjects were aware of the potential benefits they could receive
from each partner on each exchange opportunity, and thus were aware that they had one high
value (X) and one low value (Z) partner. The greater benefits potentially provided by X
combined with the fact that X also had a high value relation with Y meant subjects were more
dependent on X for resources than X was on them, and as a result, at a power disadvantage. They
were unaware of this.
8
[Insert Figure 1.]
After the instructions, subjects participated in a series of exchange opportunities with
their two simulated exchange partners. The completion of each exchange opportunity resulted in
participants receiving information about who gave them points and how many points they
received. Total earnings were updated and reported.
We divided the exchange opportunities into three phases, with the number of exchange
opportunities in each phase varying by the phase and the form of exchange. After completing
62% of exchange opportunities, the first phase stopped, and subjects were informed that they
were part way through the experiment and asked to answer a few questions about their dyadic
relationships and personal attractions for each exchange partner (Molm et al. 2007). After
answering these questions, the second exchange phase began. It stopped after a pre-determined
number of exchange opportunities, with the number varying by condition. Here subjects were
told that they had been randomly selected for a unique opportunity and that at some point in the
next part of the experiment they would be able to interact with a new set of people should they
want to stop interacting with the people with whom they had been interacting. The amount of
information subjects received about this alternative network varied by condition. Subjects then
were asked to answer questions designed to measure psychological group formation. The third
exchange phase started upon completion of the questionnaire. After a number of additional
exchanges, subjects received the opportunity to join the alternative network. The experiment
ended with the subject’s choice to stay in their current network or not.
Manipulations
Form of exchange. Participants engaged in either reciprocal or negotiated exchange.
Negotiated exchange conditions had subjects bargain with simulated partners over the division of
9
a pool of benefits. Subjects knew the range of points they could request from their partners but
did not know they were simply dividing a fixed amount of benefits when making requests,
thereby masking how many points their partner received from exchange. On any given exchange
opportunity, subjects negotiated for up to four rounds and were able to make offers, make
counteroffers, or accept offers on each round. Each round had subjects request how many points
they wanted to receive from their partners, and the computer converted these requests into offers
to the partners. Agreements occurred when an actor’s request matched her partner’s offer. If an
agreement was made, the actor immediately received that amount. If an actor failed to make an
agreement, the actor did not receive points for that exchange opportunity.
Reciprocal exchange conditions consisted of a series of exchange opportunities in which
subjects had to select to whom they wanted to give points without knowing the choices of others.
Giving points did not result in the giver losing points from his or her total. At the end of each
exchange opportunity, subjects were told who gave to them and how much they received.
Steps were taken to ensure these two forms of exchange were as similar as possible. First,
more exchange opportunities were run in reciprocal exchange conditions (160) than in negotiated
exchange conditions (80). Doubling the number of reciprocal exchange opportunities
approximately equates the time and effort spent by participants across these two forms of
exchange. Exchange opportunities were distributed across the three exchange phases with phase
1 consisting of 100 reciprocal exchange opportunities or 50 negotiated exchange opportunities,
phase 2 consisting of 40 reciprocal exchange opportunities or 20 negotiated exchange
opportunities, and phase 3 consisting of 20 reciprocal exchange opportunities or 10 negotiated
exchange opportunities.
10
Second, the monetary values of points were adjusted to ensure comparable behaviors
received comparable earnings across the two forms of exchange. The total number of points an
actor could give to an exchange partner on any exchange opportunity in reciprocal exchange was
equal to one-half the points that could be divided between actors in the negotiated exchange
setting on each opportunity.
Amount of information about the exit opportunity. We randomly placed subjects in one of
two information conditions. In the first condition, subjects were told they had an opportunity to
exchange in an alternative network but received no information about the structural dimensions
of the network. In the second condition, subjects knew they had an opportunity to exchange in an
alternative network and the only difference between their existing network and the alternative
network would be the people with whom they interacted.
Programming of Simulated Actors
We modified a program used by Molm et al. (2006) for a previous social exchange experiment to
create simulated actors. Modifications ensured simulated actors behaved in realistic ways,
appropriate to their structural positions, and were informed by behavioral data from previous
exchange experiments using only human subjects (e.g., Molm et al. 2000).
Subjects had one high power simulated exchange partner (the focal partner) and one low
power simulated exchange partner (the other partner) who each had another high power
exchange partner in common. For both forms of exchange, the behaviors of the focal partner
were set to disadvantage the low power participant, and the behaviors of the other partner were
programmed to create, on average, equal value exchanges between the low power participant and
the simulated actor. Interactions between participants and the focal partner disadvantaged the
participant so that she received, on average, 5 points for every 7 points the simulated actor
11
received from the subject. Interactions between the participant and the other simulated actor were
designed to create, on average, equal value exchanges with the participant and the simulated
partner each receiving 2 points.
Measures
Personal attraction. We measure personal attraction with four items designed to evaluate
how subjects assess each of their dyadic exchange relationships along these dimensions:
divided/united, self-oriented/team-oriented, coming apart/coming together, and
partners/adversaries (Molm et al. 2012). These measures ranged in value from one to seven, with
higher scores indicating greater attraction. Averaging item responses for the high value relation
resulted in a scale with an alpha reliability of .88. A similar scale for the low value relation had
an alpha reliability of .86.
Psychological group formation. After subjects were made aware of the alternative
network, they answered three questions designed to measure psychological group formation. The
first item had subjects indicate whether they felt very unattached/very attached to the three other
participants with whom they had been interacting (Bargozzi and Lee 2002). The second item
asked subjects to describe how much obligation (very little/very much) they felt towards these
other participants (Bargozzi and Lee 2002), and the third had subjects indicate the extent to
which they felt a sense of belonging with this set of participants (Bargozzi and Lee 2002; Thye et
al. 2011). All items ranged in value from one to seven, with higher scores indicating greater
levels of psychological group formation. The alpha reliability score for these 3 items was .84.
Staying behavior. We measured staying behavior by having participants choose to join an
alternative exchange network or not. Those who chose to stay in their existing network received
a 1. Those who chose to leave received a score of 0.
12
Exchange frequency.5 Because the frequency of exchange might vary across the two
forms of exchange and the two power conditions, we include exchange frequency as a control.
We measure the exchange frequency at each phase of the experiment by dividing the number of
agreements with (negotiated exchange) or acts of giving to (reciprocal exchange) the high value
simulated actor by the total number of exchange opportunities. Exchange frequencies ranged in
value from 0 to 1 for each exchange relation.
Results
Table 1 reports descriptive statistics by condition for the aforementioned measures. In line with
expectations, mean scores for personal attraction, psychological group formation, staying
intentions, and actual staying behaviors are generally higher for the reciprocal exchange
conditions.
[Insert Table 1.]
Although these descriptive statistics are consistent with our expectations, we test our
hypotheses more formally using a series of regression models. Our argument hinges on the idea
that the exit opportunity prompts individuals to reflect back on their past exchange relations and
that this reflection will result in disadvantaged actors being more likely to undergo psychological
group formation if those past exchanges were characterized by reciprocal, as opposed to
negotiated, exchanges. We also contend that personal attraction at the dyadic level will mediate
the relationship between the form of exchange and psychological group formation.
The regression analyses in Table 2 begin to test these claims. Model 1 regresses
psychological group formation on the form of exchange and the frequency of exchange between
the subject and focal partner. Results show only the form of exchange has a significant effect on
psychological group formation, with psychological group formation being higher in reciprocal
13
exchange conditions. Model 2 adds the two personal attraction measures. The significant main
effects for both measures show that personal attraction affects psychological group formation,
and Sobel-Goodman tests reveal that both attraction measures partially mediate the relationship
between the form of exchange and psychological group formation. These results provide support
for Hypothesis 1, 2, and 3: Following the exit opportunity, psychological group formation is
stronger for disadvantaged actors located in a reciprocal exchange network and personal
attractions partially mediate this relationship.
[Table 2.]
We also considered whether the amount of information actors have about the alternative
network affects psychological group formation. Model 3 in Table 2 reveals that it did not.
Two questions remain: Does psychological group formation affect whether individuals
ultimately choose to remain in their existing exchange networks and does it mediate the
relationship between the form of exchange and staying? Hypothesis 4 states that because
individuals who view themselves as part of a group come to favor that group over other groups,
disadvantaged actors should be more likely to choose to remain with their existing network the
more they identify as a group member. A bivariate logistic regression analysis (not shown) yields
support for this prediction.
But does psychological group formation mediate the relationship between the form of
exchange and staying in one’s network as predicted by hypothesis 6? We answer this question
using the KHB framework, which is a multistep statistical procedure used to test whether a
particular variable mediates the relationship between two variables in a nonlinear model (Karlson
and Holm 2011; Breen, Karlson, and Holm 2013).
14
Table 3 presents these results. Including psychological group formation in the model
results in a significant reduction in the main effect of the form of exchange on the choice to stay
in one’s network, as predicted by Hypothesis 5. This finding holds even after controlling for
power, information, and personal attraction.
We were surprised that including psychological group formation in the model resulted in
the coefficients for the two personal attraction measures becoming significant and negative.
Hogg (1992) contends that personal attractions are often poor predictors of behaviors at the
group level, and as a result, we did not expect there to be a significant relationship, much less a
significant negative relationship. The small sample size and the relative lack of variation in the
outcome for the negotiated exchange conditions, however, may have affected the stability of our
models and may account for this finding. We therefore ran additional models using a statistical
strategy designed for dealing with small-sample bias (Firth 1993). Results (not shown) reproduce
the findings reported using the KHB method, with the exception that the personal attraction
measures are no longer significant, as we would expect. Thus, we are confident in our
conclusions that disadvantaged actors in reciprocal exchange are more likely to stay in their
networks and that this is in part because they are more likely to undergo psychological group
formation following the exit opportunity.
EXPERIMENT 2
Our argument is that the opportunity to join an alternative network motivates psychological
group formation for those in reciprocal exchange networks but not those in negotiated exchange
networks and that this difference in group formation explains differences in staying behaviors.
While results from Experiment 1 support this, the design of Experiment 1 prevents us from
ruling rule out the possibility that differences in psychological group formation across the two
15
forms of exchange simply reflect differences in how people respond to reciprocal and negotiated
exchange. Experiment 2 attends to this possibility by examining whether psychological group
formation differs across the two forms of exchange absent the exit opportunity. If our argument
is correct, we would expect there to be no difference in psychological group formation across
these two forms of exchange.
Design and Participants
Experiment 2 was a two condition experiment that manipulated the form of exchange. 36
undergraduate students were randomly assigned to each of the two conditions. We use data from
34 of these because two students were suspicious about whether they were interacting with real
people. Each condition had 12 women and 5 men.
Procedures
Like Experiment 1, Experiment 2 involved participants engaging in a series of exchange
opportunities with two exchange partners in what they believed was a fixed, four-person
exchange network. The procedures were similar to those used in Experiment 1. The major
difference was that after a set number exchange opportunities, Experiment 2 simply ended and
participants were asked to fill out a questionnaire evaluating the extent to which they identified
with the network as a group.
Similar to Experiment 1, there were more exchange opportunities in reciprocal exchange
conditions (100) than in negotiated exchange conditions (50) and monetary values of points were
adjusted to ensure comparable behaviors received comparable earnings across the two forms of
exchange.
Measures
16
Psychological group formation. Our measure of psychological group formation captures
whether participants felt 1. attached to the other members of the network, 2. a sense of obligation
to the other individuals, and 3. a feeling of belonging with the other individuals in the network.
The alpha reliability for these three items was .77.
Exchange frequency. The primary control variable for this analysis was the frequency of
exchange. We measured this in the same way as for experiment 1, and values ranged from 0 to 1
for each exchange relation.
Results
To see if the form of exchange affected psychological group formation absent the exit
opportunity, we regressed psychological group formation on the form of exchange controlling
for exchange frequency. Neither variable was significant at the .05 level. This finding
strengthens our argument that it is the exit opportunity that introduces into the situation a
categorical distinction that under the right conditions, serves as the basis for disadvantage actors’
psychological group formation.
DISCUSSION
Our research examines whether a history of reciprocal exchange makes disadvantaged actors
more likely to stay in exchange networks than a history of negotiated exchange. It shows that the
greater personal attractions produced by reciprocal exchange prime disadvantaged actors such
that they are more likely to view themselves as group members following an exit opportunity.
Thus, our research points to the exit opportunity as one factor that can motivate psychological
group formation for disadvantaged actors so long as it is primed by an appropriate exchange
history. This is important, because group identification mediates a significant portion of the
effect of exchange form on disadvantaged actors’ staying behaviors.
17
This finding speaks to an apparent contradiction in the social exchange literature. While
Emerson (1972) argues that power imbalances in exchange processes should produce pressures
toward structural change through coalition formation and network expansion, Blau (1964)
suggests that power imbalances could become institutionalized and resistant to change. We
account for these contradictory assertions by showing how the form of exchange can influence
the likelihood of an alternative structural change mechanism: exit. Unlike coalition formation
and network expansion, which involve simply modifying linkages within a network structure,
exit involves severing ties with all members of a network and thus, is another factor to consider
when investigating how people respond to structural disadvantage.
Our findings also cast a shadow on the value of reciprocity. While Molm and colleagues
convincingly demonstrate that integrative bonds between exchange partners are stronger in pure
reciprocal exchange than in pure negotiated exchange (e.g., Molm et al. 2007; Molm et al. 2012),
our research reveals that this can result in disadvantaged actors foregoing opportunities to
improve their situation. Exactly because reciprocity veils issues of fairness by reducing conflict
(Molm et al. 2003) and heightens the expressive value of exchange (Molm et al. 2007), it
increases the likelihood of disadvantaged actors adopting an attachment to the network that can
retard structural change. Thus, reciprocity as an integrating and stabilizing force may increase
the likelihood of actors making choices that reinforce their disadvantage.
By addressing issues of stability and inequality concurrently, this research yields valuable
insights that have potentially broad implications. For instance, our research suggests turnover
rates should be higher in organizations characterized by a higher frequency of negotiated
interaction, perhaps because this form of exchange yields a weaker psychological contract
(Rousseau 1989). Similarly, our research may help explain the persistence or dissolution of
18
marriages and families. Collett and Avelis (2011) reveal that exchange patterns in families tend
to be dominated by either reciprocity or negotiation. We, therefore, can surmise that low power
family members will be more likely to sever ties in families characterized by negotiation. While
these hypotheses are mere speculations, they follow from the logic of our theory and point to its
general applicability. Thus, by explaining how micro-level processes can affect choices that
ultimately affect the stability of macro-level structures, our research illustrates how theories of
social exchange can inform sociological research more generally.
19
SUPPLEMENTAL MATERIAL
Additional supporting information may be found at spq.sagepub.com/supplemental.
20
ACKNOWLEDGEMENTS
We would like to thank Linda Molm, Joseph Galaskiewicz, Kelly Bergstrand, David Melamed,
and the members of the Social Psychology Seminar at University of California, Riverside for
their advice on this paper. Earlier versions of this paper were presented at the 22nd Annual Group
Processes Conference in Atlanta, Georgia 2010 and the 110th Annual Meeting of the American
Sociological Association in Chicago, Illinois.
21
FUNDING
This project was funded by a National Science Foundation Dissertation Improvement Grant
(SES-1003281) awarded to Linda Molm and Scott V. Savage.
22
NOTES
1. In addition to the form of exchange and information, Experiment 1 manipulated power by
placing participants in either high or low power conditions. While the manuscript focuses on the
low power conditions, online Appendix A reports the results of a mediation analysis using the
KHB method to examine the effects of power. This analysis reveals that power moderates the
effects of psychological group formation on staying behavior; the effects are strongest for those
in low power conditions. This finding strengthens our confidence in our argument.
2. 67 undergraduate students actually participated in these conditions. We excluded data from
three subjects, because they did not believe they were interacting with real people.
3. Sensitivity analyses found no gender effects.
4. Subjects exchanged in negatively connected networks (Cook and Emerson 1978). To
operationalize a negatively connected network under conditions of negotiated exchange, actors
could make an agreement with one, and only one, exchange partner on any given exchange
opportunity. To do so in reciprocal exchange conditions, actors could only give to one exchange
partner on any given exchange opportunity.
5. Our measure of exchange frequency was relatively highly correlated with earnings (.77). We,
therefore, did not include earnings in the statistical models we present. Doing so, however, does
not result in different conclusions. See online Appendix B for these results.
23
REFERENCES
Adler, Patricia A. and Peter Adler. 1995. “Dynamics of Inclusion and Exclusion in Preadolescent
Cliques.” Social Psychology Quarterly 58(3):145-162.
Backman, Carl W. and Paul F. Secord. 1962. “Liking, Selective Interaction, and Misperception
in Congruent Interpersonal Relations.” Sociometry 25(4):321-335.
Bargozzi, Richard P. and Kyu-Hyun Lee. 2002. “Multiple Routes for Social Influence: The Role
of Compliance, Internalization, and Social Identity.” Social Psychology Quarterly
65(3):226-247.
Blau, Peter M. 1964. Exchange and Power in Social Life. New York: Wiley.
Breen, Richard, Kristian Bernt Karlson, and Anders Holm. 2013. “Total, Direct, and Indirect
Effects in Logit and Probit Models.” Sociological Methods & Research 42(2):164–91.
Cheshire, Coye, Alexandra Gerbasi, and Karen S. Cook. 2010. “Trust and Transitions in Modes
of Exchange.” Social Psychology Quarterly 73(2):176-195.
Collett, Jessica L., and Jade Avelis. 2011. “Building a Life Together: Reciprocal and Negotiated
Exchange in Fragile Families.” Pp.227-254 in Advances in Group Processes, vol 28,
edited by Shane R. Thye and Edward J. Lawler. United Kingdom: Emerald Group
Publishing Limited.
Cook, Karen S. and Richard M. Emerson. 1978. “Power, Equity and Commitment in Exchange
Networks.” American Sociological Review 43(5):721-39.
Dogan, Gonul, Macel A.L.M van Assen, Arnout van de Rijt, and Vincent Buskens. 2009. “The
Stability of Exchange Networks.” Social Networks 31(2):118-125.
24
Emerson, Richard M. 1972. “Exchange Theory, Part II: Exchange Relations and Networks.”
Pp.58-87 in Sociological Theories in Progress, vol 2, edited by J. Berger, M. Zelditch,
Jr., and B. Anderson. Boston: Houghton-Mifflin.
Firth, David. 1993. “Bias Reduction of Maximum Likelihood Estimates.” Biometrika 80(1):2738.
Hirschman, Albert O. 1970. Exit, Voice, and Loyalty: Responses to Decline in Firms,
Organizations, and States. Cambridge, MA: Harvard University Press.
Hogg, Michael A. 1992. The Social Psychology of Group Cohesiveness: From Attraction to
Social Identity. New York: New York University Press.
Hogg, Michael A. and John C. Turner. 1985. “Interpersonal Attraction, Social Identification, and
Psychological Group Formation.” European Journal of Social Psychology 15(1):51-66.
Homans, George C. 1974. Social Behavior: Its Elementary Forms, rev. ed. New York: Harcourt
Brace Jovanovich.
Karlson, Kristian Bernt and Anders Holm. 2011. “Decomposing Primary and Secondary Effects:
A New Decomposition Method.” Research in Social Stratification and Mobility
29(2):221–37.
Kmec, Julie A. 2007. “Ties that Bind? Race and Networks in Job Turnover.” Social Problems
54(4):483-503.
Kurtzberg, Terri and Victorica Husted Medvec. 1999. “Can We Negotiate and Still Be Friends?”
Negotiation Journal 15(4):355-361.
Kuwabara, Ko. 2011. “Cohesion, Cooperation, and the Value of Doing Things Together: How
Economic Exchange Creates Relational Bonds.” American Sociological Review
76(4):560-580.
25
Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: Test of a
Theory of Relational Cohesion.” American Sociological Review 61(1):89-108.
Lawler, Edward J., Shane R. Thye, and Jeongkoo Yoon. 2008. “Social Exchange and Micro
Social Order.” American Sociological Review 73(4):519-542.
_____. 2009. Social Commitments in a Depersonalized World. New York: Russell Sage
Foundation.
Leik, Robert K. 1992. “New Directions for Network Exchange Theory: Strategic Manipulation
of Network Linkages.” Social Networks 14(3-4):309-323.
Molm, Linda D. 1994. “Dependence and Risk: Transforming the Structure of Social Exchange.”
Social Psychology Quarterly 57(3):163-176.
Molm, Linda D. and Karen S. Cook. 1995. “Social Exchange and Exchange Networks.” Pp.209235 in Sociological Perspectives on Social Psychology. Edited by K. S. Cook, G. A. Fine,
and J. S. House. Boston: Allyn Bacon.
Molm, Linda D., Jessica L. Collett, and David R. Schaefer. 2006. “Conflict and Fairness in
Social Exchange.” Social Forces 84(4):2331-52.
Molm, Linda D., Jessica L. Collett, and David R. Schaefer. 2007. “Building Solidarity through
Generalized Exchange: A Theory of Reciprocity.” American Journal of Sociology
113(1):205-42.
Molm, Linda D., David R. Schaefer, and Jessica L. Collett. 2009. “Fragile and Resilient Trust:
Risk and Uncertainty in Negotiated and Reciprocal Exchange.” Sociological Theory
27(1):1-32.
26
Molm, Linda D., Nobuyuki Takahashi, and Gretchen Peterson. 2000. “Risk and Trust in Social
Exchange: An Experimental Test of a Classical Proposition.” American Journal of
Sociology 105(5):1396-1427.
Molm, Linda D., Nobuyuki Takahashi, and Gretchen Peterson. 2003. “In the Eye of the
Beholder: Procedural Justice in Social Exchange.” American Sociological Review
68(1):128-152.
Molm, Linda D., Monica M. Whitham, and David Melamed. 2012. “Forms of Exchange and
Integrative Bonds: Effects of History and Embeddedness.” American Sociological Review
77(1):141-165.
Rand, David G., Samuel Arbesman, and Nicholas A. Christakis. 2011. “Dynamic Social
Networks Promote Cooperation in Experiments with Humans.” Proceedings of the
National Academy of Sciences 108(48):19193-19198.
Rousseau, Denise M. 1989. "Psychological and Implied Contracts in Organizations." Employee
Rights and Responsibilities Journal 2:121-139.
Rousseau, Denise M., and Snehal A. Tijoriwala. 1998. "Assessing Psychological Contracts:
Issues, Alternatives, and Measures." Journal of Organizational Behavior 19:679-695.
Thye, Shane R., Edward J. Lawler, and Jeongkoo Yoon. 2011. “The Emergence of Embedded
Relations and Group Formation in Networks of Competition.” Social Psychology
Quarterly 74(4): 387-413.
Turner, John C., and Katherine J. Reynolds. 2012. “Self-Categorization Theory.” Pp.399-417 in
Handbooks of Theories of Social Psychology: Volume 2. edited by P. A.M. Van Lange,
A. W. Kruglanski, and E. T. Higgins. Thousand Oaks, CA: Sage Publications Ltd.
27
Turner, John C., Penelope J. Oakes, S. Alexander Haslam, and Craig McGarty. 1994. “Self and
Collective: Cognition and Social Context.” Personality and Social Psychology Bulletin
20(5):454-463.
Willer, Robb and David Willer. 2000. “Exploring Dynamic Networks: Hypotheses and
Conjectures.” Social Networks 22(3):251-272.
28
AUTHOR BIOS
Scott V. Savage is an assistant professor in the Department of Sociology at the University of
Houston. Currently, with support from the National Science Foundation, he and his collaborators
are investigating whether person identities can overcome structural pressures to exploit others, as
well as how identities can generate social inequality when structural pressures are absent. He is
also studying how status processes affect work group dynamics. His recent research appears in
The Sociological Quarterly (2016) and American Journal of Sociology (forthcoming).
Zachary Sommer received his BA in Sociology at Augsburg college in 2010, and his MA in
Sociology at the University of California, Riverside in 2012. His primary research interests focus
on the role of human emotions and identities in exchange processes.
29
TABLES
Table 1 Means and Standard Deviations for Measures by Experimental Condition
(N=64)
Reciprocal Exchange
Negotiated Exchange
Measures
Personal
Attractions For
Focal Actor
Low Info
High Info
Low Info
High Info
4.94
(1.04)
4.48
(1.22)
4.00
(1.56)
4.17
(.70)
Other Actor
3.28
(1.21)
3.75
(1.19)
3.03
(1.33)
3.05
(1.27)
Psychological
Group
Formation
4.33
(1.32)
4.54
(1.04)
3.5
(1.22)
2.88
(1.41)
.5
.5
.19
0
.67
(.15)
.70
(.14)
16
16
Staying
Behavior
Exchange
.76
.67
Frequency
(.12)
(.15)
with Focal
Actor
N
16
16
Note: Standard deviations in parentheses.
30
Table 2 Unstandardized OLS Coefficients from the Regression of Psychological
Group Formation on Independent Variables (N = 64)
Independent Variable
Model 1
Model 2
Model 3
1.26*
Form of Exchange (Rec. =1; Neg. =0)
.81*
.81*
(.32)
(.32)
(.32)
Exchange Frequency with Focal
Partner
-.24
(1.13)
-.29
(1.09)
-.34
(1.09)
.50*
(.14)
.29*
(.13)
.50*
(.14)
.29*
(.13)
-.22
(.29)
3.35*
(.80)
.48
(1.13)
.61
(1.15)
.21
.36
.37
Personal Attraction with Focal Partner
Personal Attraction with Other Partner
Information
Constant
R2
* p < .05 (two-tailed tests)
Note: Standard errors are in parentheses.
31
Table 3 Logistic Regression Coefficients Predicting Staying in One’s Social
Network. The change in coefficients was derived using the KHB method. (N =64)
Independent
Model 1a
Model 2
Change in β
Variable
Form of Exchange
(Rec. = 1; Neg. = 0)
4.36*
(1.31)
2.48*
(1.00)
Exchange
Frequency with
Focal Partner
-.93
(3.18)
-.14
(3.16)
Personal Attraction
for Focal Actor
-.10
(.42)
-1.26*
(.58)
Personal Attraction
for Other Actor
-.32
(.41)
-1.00*
(.48)
Information
-1.70
(.99)
-1.19
(.93)
Constant
-1.36
(3.49)
-2.78
(3.51)
Psychological
Group Formation
1.88*
(.90)
2.32*
(.69)
Note: Standard errors are in parenthese.
* p < .05 (two-tailed test)
a
The coefficients from this model differ from a standard logistic regression model
because the residualized psychological group formation measure is included in the
model, thereby making the coefficients in this model comparable to those in Model 2.
We do not report the coefficient for the residual here.
32
FIGURE
Figure 1. Diagram of the network that appeared on subjects’ screens. Subjects were told that they
were ‘W’ and that they could exchange with ‘X’ and ‘Z’.
33